基于细节信息分类和结构相似的快速分形编码方法  

Fast fractal coding based on detail information classification and structure similarity

作　　者：孟帆[1] 叶正麟[1] 潘璐璐[1] 

机构地区：[1]西北工业大学理学院,陕西西安710072

出　　处：《计算机应用》2005年第12期2849-2852,2864,共5页journal of Computer Applications

基　　金：国家自然科学基金资助项目(10070006)

摘　　要：针对传统分形编码中编码时间过长的问题,提出了一种细节信息分类(DetailInformationClassification,DIC)与结构相似(StructureSimilarity,SSIM)指标相结合的快速分形编码方法。DIC方法将图像块的内部细节信息分布状况作为分类标准对定义域块库进行分类,SSIM指标分别从亮度、对比度和结构三个方面度量值域块与定义域块之间的匹配程度。通过将DIC方法与SSIM指标结合,可以加快分类速度、减少匹配运算时间。The conventional fractal coding scheme suffers from long encoding time. A new fractal coding method was proposed based on detail information classification（DIC） and structure similarity（SSIM） index. DIC was implemented based on distribution of detail information which was discrepancy between pixel and its adjacent pixels. SSIM index measured the similarity between range block and domain block from luminance, contrast and structure. The combination of DIC and SSIM index can accelerate the classification and reduce time of matching computation. Experiment results on test images show that the presented method can decrease the coding time whilst maintain decoding quality.

关 键 词：分形编码 细节信息分类 结构相似指标 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]